optimizing program return release final1optimizing your loyalty program return maximize the...

Loyalty Improvement Series Copyright © Paytronix Systems, Inc. 2010. All rights reserved.

www.paytronix.com

Optimizing Your Loyalty Program Return

WHITE PAPER Optimizing Your Loyalty Program Return Maximize the profitability of your loyalty program by finding your program’s ideal reward-value range.

Overview

A key aspect of designing and managing a loyalty program is setting its reward

values. Determining how much to give back to your customers as recognition for

exhibiting revenue-generating behaviors, such as spending more at your

restaurant or visiting more frequently, takes careful consideration. Giving too little

does not motivate a change in behavior, while giving too much may unnecessarily

erode profits.

Many of our customers have asked if there is an optimal setting for a loyalty

program’s rewards structure. Intuitively speaking, the higher the value your guests

place on your reward program, the higher their propensity to join and the more

often they will visit. However, does the cost of an increasingly rich rewards

structure affect the program’s overall return? In this Loyalty Improvement Series

white paper, we will examine the factors involved in determining program reward

levels

and explain how to map out an organization’s profitable reward zone. Whether you

have an established program or are thinking about launching one, understanding

the zone in which your program is likely to operate at an optimized profitability

level will help you make the decisions that impact the success of your loyalty

program.

Program Return is the amount of revenue generated from your loyalty program, minus associated program costs.

Variable Cost Rate is expressed as a percentage of incremental revenue generated by a loyalty program. It accounts for expenses such as food and labor costs, reward costs, and program administrative costs.

Figure 1A: A sample optimization curve for a specific variable food and labor cost percentage.

Results Applicability The results of this analysis were derived from a set of quick casual clients with similar check averages. While we expect similar findings for casual and fine dining, the detail results may vary by concept type and check average.


2


There are several factors that contribute to program return, including perceived

reward value, buy-more rate, visit rate, member acquisition and costs. We created

a model that incorporates these influencing factors, basing it on Paytronix client

data. With this model, we

can input a variable cost

rate to derive a reward-

optimization curve that

predicts the impact that

changes in reward levels

will have on a program’s

return. The reward-

optimization curve,

generated with our

model, uncovers a

reward zone showing

where a loyalty program

designer should set reward levels to maximize the program benefit. We

recommend targeting a zone that promises between 80% and 100% of the

program benefit for your core program, while leaving room to enrich your offer

during short-term promotions. Figure 1B summarizes the reward profitability zones

(between 80% and 100% of the maximum program return) for a variety of variable

cost rates. The range is attractive because of the shape of the optimization curve

itself. Since the curve is flat around its peak, a variance between a concept’s

loyalty program and this model will not significantly compromise the program’s

return.

This paper goes into detail about each element of the reward-optimization model.

We share the methodology used to create the model and answer the following

questions:

1. What are the relationships between reward level and visits, spend, and

membership?

2. How do those relationships impact a program’s profitability?

3. Where should a reward structure be set, given a restaurant’s variable food

and labor cost?

Our Methodology

Accessing a broad data set was imperative to this study. Based on our client data,

we focused on quick-service concepts so that the data was derived from a

relatively homogeneous group of concepts with similar check sizes and reward

styles (automatic rewards). In addition, we were only interested in merchants who

Figure 1B

The Perceived Value of Rewards

How do guests value the rewards you

offer within your program?

This is an important question to

answer prior to establishing your

reward types. In this analysis, we

assume that guests’ perception of

reward value is equal to the reward’s

retail value. For example, if a guest is

given a $10 off coupon or an entrée

item that also has a retail value of

$10, we assume, for our analysis,

that the guest values both at $10.

Here is an example of how we

calculated the perceived value of a

sample rewards program.

Program Structure:

Buy 10 Burritos, Get 1 Free

Assumptions:

Guests typically buy one burrito and

one drink at each visit.

The retail value of a burrito is $6.50.

The retail value of a drink is $1.50.

The total average check in this scenario

is $8.00.

It takes 10 visits or a total spend of

$80 {10 X ($6.50+$1.50)} to reach

the “free burrito” reward. In this case,

the perceived value of the reward, or

give-back rate, is calculated as

follows:

Perceived Value =

= $ .

$ = 8.125%


3


had been operating loyalty programs for at least three years.

We began creating the reward-optimization model by first making some

observations regarding loyalty-member behavior and how it relates to the

perceived value of the program benefits. This resulted in three basic principles

being observed in data from our clients’ loyalty programs.

As the perceived value of a program’s rewards increases, members will:

1. Buy more per visit.

2. Visit more frequently.

3. Join at a faster rate.

The corollary of these principles is that each has a unique impact on a program’s

return. Figure 2 illustrates how the perceived value of rewards drives the buy-more

rate, the visit rate, and member acquisition, all of which contribute to the ROI

calculation in our model and, thereby to the program’s overall return. (For a further

explanation on the ROI model, please see the Loyalty Improvement Series article

“Building Your Loyalty Program ROI.”)

Figure 2

Reward Cost in this model is the food and labor cost of the rewards. This excludes costs such as rent, utilities, and interest, which should not increase based on giving a reward.

Member Acquisition is a measure of the number of active members per restaurant who are participating in a given program.

Visit Rate is the percentage increase in guest visit frequency due to the loyalty program.

Buy-More Rate is the percentage increase in member spending due to the loyalty program.


4


Our analysis determined how spend, visits, and membership varied by the level of

rewards offered and how each relationship impacts incremental revenue and

incremental costs. We derived three mathematical functions, or curves, for these

relationships and approached the establishment of each curve in the same

manner. First, based on our experience, we hypothesized the shape of the curve

to establish an estimated curve. Second, we graphed concrete data points from

similar client programs. Last, we adjusted our estimated curve to reflect the actual

data so that our curve would be an accurate representation of reality.

The Relationship Between Reward Level, Buy-More Rate, Visit Rate, and Member Acquisition

Buy-More Rate

Going a step further with the guest-behavior principle that members buy more as

reward levels increase, our expectation was that the curve representing the impact

of increased reward value would follow the law of diminishing returns. When

guests redeem their rewards, the reward is subtracted from their check total, so we

subtract the cost of the rewards when calculating the increase in spend (see the

middle graph in Figure 3).1 The resulting hypothesized shape of the buy-more

curve is shown in the graph to the right in Figure 3.

As you can see, raising reward levels increases buying behavior to a certain point,

and then spending begins to diminish. Aggregating the data from several loyalty

programs enabled us to model the likely difference in average check between a

nonmember and a member as a function of reward levels. To establish our curve,

we compared the check averages of member to nonmember for our quick casual

1 It is important to note that Paytronix measures spending after rewards have been redeemed from the check. Because the measurement is “net of rewards,” we need to deduct the cost of rewards from any projected increase in spend that would come from the motivational incentive that the reward structure creates for the guests to spend more.

Figure 3

Promotional Programs are tools used to engage members through the use of ad hoc offers that are designed to further compel specific desired behaviors.

Layered Programs add depth to your core loyalty program. Program layers can take several forms, including: giving a donation to a worthy cause for every member visit, creating a unique birthday or anniversary program, and adding a surprise-and-delight reward scheme.

Core Program is the heart of your loyalty program, that attracts guests to join and compels them to identify themselves at each visit.

Loyalty Program Building

Blocks


5


clients who have similar-size check averages. In each case, we found that the

check average among loyalty members was more than the overall check average.

The results were plotted against the perceived value of each customer’s rewards

program and a best-match curve was established to fit the data points, resulting in

the buy-more curve shown in Figure 4. Taking a closer look, we made four distinct

observations from the data.

Observation 1: Client A’s position on the graph shows that there is a minimum

reward level that must be exceeded before a program has an impact on guest

behavior. With a 3% perceived reward value, average guest checks among

members were nearly identical to the overall average check for the restaurant.

Members in this program are not being motivated to spend more per visit.

Observation 2: Client C has nearly maximized the benefit of its rewards’

perceived value. With about a 10% perceived reward value, this customer is

motivating its members to spend greater than 15% more than its average guest.

Adding to the reward value will likely not increase the average check, as

demonstrated by the shape of the curve.

Observation 3: Client D and E could afford to scale back the perceived value of

their rewards while maintaining a 12-15% lift in average check among program

members.

Observation 4: Client F (the point of data that appears far above our curve) can

be explained in a couple of ways. After plotting this data set, our first question was,

“What are they doing right?” After all, our clients want to offer rewards at the

lowest cost to the business while compelling high involvement in their loyalty

Minimum Reward Level

There is a reward threshold required to affect guest behavior. The results of our analysis show that the minimum level is somewhere between 0% and 5%. If you run a rewards program below a 5% reward value, it is unlikely that you will motivate your members to buy more or visit more.

Do guests value all rewards

equally?

Perceived value for a specific menu

item or piece of merchandise is a bit

more difficult to ascertain when

compared to a more general reward,

such as a $10 dining certificate or a

complimentary menu category item

(e.g., an entrée).

If you limit the reward to a coffee, for

example, not all guests value a cup

of coffee at the same rate. A portion

of your guests will have no interest in

coffee and therefore assign a

perceived value of $0 to the reward.

In some cases, if a guest receives

the same irrelevant reward

repeatedly, the guest may actually

associate a negative value to this

reward and feel that you do not

respect his or her preferences.

On the other hand, if your guest has

a passion for coffee, that guest may

place a perceived value on the

coffee reward that exceeds its retail

value.

Figure 4


6


programs. We believe this restaurant, Client F, is enjoying higher-than-normal

average checks among its members because it has succeeded in creating a high

perceived value for its rewards.

Visit-More Rate

Measuring a change in visit behavior in the absence of a loyalty program is difficult

for two reasons. First, since restaurants do not have the capability to track the

purchasing behavior of guests before the existence of a loyalty program there is an

absence of data with which to compare post-program results. Second, as soon as

a loyalty program is introduced, guest behavior instantly changes. For example,

when a new loyalty program is rolled out, some guests will immediately switch their

behavior from visiting several different restaurants for lunch during the week to

visiting the restaurant with the new loyalty program multiple times per week.

With each visit, the guest may be taking lunch orders for colleagues so that more

loyalty program points can be earned. In this scenario, the guest’s purchasing and

visiting behavior prior to the loyalty program was drastically different than it was

after the program’s launch. Guest behavior changes quickly upon implementation

of a new loyalty program; therefore, guest behavior measured during the early

phase of a program is not a good surrogate for pre-program behavior.

Despite the lack of pre-program data, the visit-rate curve can be derived by

analyzing the change in visit rate caused by limited-time promotions. In lieu of

comparing a “pre-program visit rate” against a “post program visit rate,” we

analyze the incremental visits that occur during promotions. The likely change in

visits that would occur with increasing reward levels can be extrapolated from data

representing the short-term impact on visits resulting from an increase in reward

value during promotions.

As members get closer to a

reward, they will make a greater

effort to achieve the goal.

In an article published in the Journal

of Marketing Research, authors

Kivetz, Urminsky, and Zheng report

their findings in member behavior

from two empirical tests. They found

that “members of a café RP (e.g.,

“buy ten coffees, get one free

[loyalty program]”) purchase coffee

more frequently the closer they are

to earning a free coffee (on average,

inter-purchase times decrease by

20% or .7 days throughout the

program).”

Further findings include, “members

were more likely to defect when they

were farther away from the reward

goal (or, equivalently, they were

more likely to persist when they

were closer to the goal.”

Source: Ran Kivetz, Oleg Urminsky,

and Yuhuang Zheng, “The Goal-Gradient Hypothesis

Resurrected: Purchase Acceleration, Illusionary Goal

Progress, and Customer Retention,” Journal of Marketing Research,

Feb. 2006


7


We predicted a curve that represented the impact that reward value has on

member visits. From our experience, we know that an increase in the reward value

should increase visits, and as the member gets closer to receiving a reward, visit

rate should accelerate. We also know that the change in visit rate for different

types of promotions varies. For example, a limited-time offer (LTO) for a “free

cookie with visit” will increase the perceived reward value slightly, thereby

increasing the visit rate slightly (see triangle C in Figure 5). Offering double points

for visits, on the other hand, will increase the change in visits considerably (see

triangle F in Figure 5). Triple-point offers will compel a greater increase in visits

than a double-point offer, but the incremental benefit is less. Measuring the slope

at different points along the visit-more curve during promotion periods provided the

information needed to calculate the increase in visit rate due to the core loyalty

program.

The formula we used to graph the visit-rate curve is:

∗ √

Since we can only measure slopes of this curve, we took its derivative, as shown

below:

√

∗

√

After plotting our estimated curve, we analyzed the results of several well-defined

client promotions, including double-point offers, limited-time offers, bonus-point

Figure 5

Figure 5 Key

Client 1: Free Cookie LTO A: visits at original reward- value level B: visits with increase in reward-

value level for “free cookie” LTO C: change in visit rate as a result of the change in the reward value Client 2: Double Points Offer D: visits at original reward- value level E: visits with increase in reward-

value level for double points F: change in visit rate as a result of the change in the reward value

Increasing the Perceived Value

of Rewards

There are two ways to increase the

perceived value of your rewards:

1. Use highly relevant rewards

to motivate guest behavior. Rewards should appeal to

guests based on their specific

preferences. If they drink

coffee, their reward is coffee,

while if other guests favors

donuts, their reward may be an

extra donut.

2. Compel spending by offering

exclusive or limited-supply

rewards. In a fine-dining

environment, the reward could

be an exclusive invitation to a

special wine tasting, a chef’s

dinner, or a unique excursion

for which only a certain number

of seats are available.


8


offers, and visit challenges. Because we were interested in seeing the change in

visit rate, we studied the “before promotion period” and compared it with the

“during promotion period,” which uncovered a clear change in visit behavior. The

data closely replicated the shape of our estimated curve (see Figure 6A). We

adjusted our estimated curve slightly to fit the line you see in Figure 6B.

Member Acquisition

The number of members you attract to your program is directly related to the value

of your reward structure. Higher-value programs enroll more members who will

actively participate in the program. As we did with the buy-more and visit-more

curves, we assume that the law of diminishing returns applies to the relationship

between the reward-value and membership levels.

Several Paytronix clients have changed program reward structures during the life

of their programs. As the perceived value of the rewards change, we see

merchants experience changes in enrollment activity and ongoing member

participation. We can clearly see this effect by isolating one particular client that

changed its reward structure three different times. This unique data set allows us

to see the direct impact that reward value has on member behavior. In addition,

the reward values within the data set cover the most interesting reward-level

range, 5% to 20%.

Figure 6A Figure 6B

Free or Fee?

The programs analyzed within this

paper either were free to join or

had a nominal fee ($1 or less)

associated with them.

Whether or not you charge for

membership to your loyalty

program is a key decision. Free

membership will engage a larger

guest base and can motivate broad

sales gains. Charging a fee will

attract committed guests and

speed time to profitability.

See the Loyalty Improvement

Series article “Free or Fee: Should

You Charge for Your Loyalty

Program?” for more details.


9


Figure 7 graphically depicts the impact that the three different reward levels had on member participation per location. The program began with a reward rate of 17.9% (point A), which attracted the highest active membership level over the program’s life. When the reward level was cut back to 5%, there was a dramatic decrease in the number of active members, as shown by point B in Figure 7. The client then increased the program reward level to 10% and experienced a significant increase in member activity (point C).

We used this membership activity data to establish a best-fit member-acquisition curve, which is overlaid onto this graph and used in the optimization model.

Deriving the Return-Optimization Curves

Once we established the buy-more, visit-more, and member-acquisition curves, we

plugged them into our return-optimization model to find out what the best reward

levels are, given an organization’s variable cost structure.

Our model includes related calculations for the incremental revenue that occurs

from members who visit and spend more than nonmembers, along with the

associated incremental food and labor costs. In addition, we included the program

reward costs by multiplying the member-generated revenue by the retail value of

the reward and the variable cost rate. Program costs also come into play when

considering a program’s return on investment. Data from multiple client programs

depicted a median program cost rate that was 1.6% of the incremental revenue

generated by the loyalty program. We then plotted the results of the model based

on several levels of variable cost rates, from 40% to 60%. See curves in Figure 8.

The final set of curves represents the relationship of all loyalty program elements,

Figure 7

Double Points or Triple Points?

Most recommended core-reward

levels enable you to offer a double-

point promotion and still receive at

least 80% of the value of your

program.

Triple-point promotions increase

the value of the rewards beyond

80% optimization. In certain

instances, aggressive promotions

may make sense.

Temporarily reducing the program’s

return may produce long-term

program benefits. For example,

running a short-term triple-point

promotion with the purpose of

igniting enrollment means you will

have more members in your

program who will spend more at

each visit and will visit more


10


beginning with the effect that reward levels have on spending, visit behavior, and

membership levels. As reward levels increase (x axis), program return increases at

different rates depending on the restaurant’s variable cost rate. The lower the

variable costs, the higher the return will be, up to a certain reward level. For

example, if a restaurant’s variable cost rate is 60%, the return for the program

begins to diminish after about a 9.5% reward value. Conversely, when the variable

cost rate is 40%, the point of diminishing returns begins after a reward level of

nearly 18%. Higher reward levels mean more members, which in turn translates to

higher program returns.

Put the Loyalty Program Optimization Curves to Use in Your Restaurant

When the interrelationship between reward value, member behavior, and program

costs is graphed, the curve reveals a range of reward values that are optimal for

the program.

The return-optimization curves in Figure 8 have a shallow slope around their

peaks. At the peak, or optimal return point, you would reap 100% of the program’s

potential return. However, there is a point to the left of the peak where you would

capture 80% of the program’s value and another point to the right of the peak

where you would enjoy 80% of the program’s value. We have defined the area

between these two 80% points to be the zone within which program rewards

should be set. Your core-program reward level should be set at any point where

reward levels bring from 80% to 100% of the value of the program. When extra

rewards are added to your program for layers or promotions, set the reward value

Figure 8

Loyalty Program Optimization Curves

Buy 10, Get 1 Free

Interestingly, over the years, most

restaurateurs have found reward

levels that work for their individual

organizations. Whether they

employed scientific testing or trial

and error to find a level that works

best for them, most reward

programs out there today offer

reward levels at about a 10% give-

back rate, such as a “buy 10, get 1

free” program.

A 10% reward value strikes the

right balance between giving

enough to motivate guest behavior

and profitability. Typically, 10% is

within the profitability zone for

restaurants with variable costs

between 40% and 55% of sales.


11


between the 100% optimization point and a higher reward value so that you reap

at least 80% of the program’s profitability. Leaving room for additional promotions

will keep your guests excited and compel them to be continuously engaged in your

program. Setting your reward level outside of this zone will result in lower program

profitability levels.

Conclusion

Setting reward levels is not an exact science. Understanding the intricate factors

that impact a program’s return will enable you to determine the reward level that

makes sense for your organization. It is important to understand what range of

optimization is acceptable to you. Once you have identified this range, be sure to

set your core-program reward value at a level that allows you to periodically enrich

your program in order to maintain a high level of member involvement.


12


About the Author

Andrew Robbins, President, Paytronix Systems, Inc.

Andrew concentrates on product development and design. He is the visionary behind the Paytronix Solution, making sure that the company stays several steps ahead of its competitors.

Prior to establishing Paytronix, Andrew co-founded Sparkventures, LLC, a venture capital firm specializing in software technology. He also worked for Caradon Doors and Windows, where he was Vice President of Engineering. Caradon supplies products to home improvement outlets, including Home Depot and Lowe’s.

Andrew previously spent seven years as an engineer with General Electric, designing computer controls for jet engines.

Andrew has a BS in mechanical engineering from Princeton University, an MS in mechanical engineering from Massachusetts Institute of Technology, and an MBA from the Harvard Business School. He lives in Newton, MA, with his wife and two daughters.

About Paytronix

Paytronix is a leading software‐as‐a‐service provider of gift, loyalty, and email solutions. Its unique platform is the most versatile in the industry, enabling restaurants to drive measureable revenue increases. Through its innovative software design and integrations with widely used restaurant POS systems, Paytronix empowers its clients with the greatest amount of flexibility to build unique, brand-enhancing programs.

Paytronix Systems, Inc. 307 Waverley Oaks Road, Suite 309

Waltham, MA 02452 617.649.3300

[email protected] www.paytronix.com

optimizing program return release final1optimizing your loyalty program return maximize the...

Documents